Farm cooperatives, household vulnerability
and agricultural child labour in Rwanda
UCW Country Report July 2016
© International Labour Organization/ Crozet M.
Farm cooperatives, household vulnerability and agricultural child labour in Rwanda
UCW Country report
July 2016
Understanding Children’s Work (UCW) Programme
International Labour Organization ILO Office for Italy and San Marino
Villa Aldobrandini V. Panisperna 28 00184 Rome
Tel.: +39 06.4341.2008 Fax: +39 06.6792.197
Centre for Economic and International Studies (CEIS)
University of Rome ‘Tor Vergata’ Via Columbia 2 - 00133 Rome
Tel.: +39 0672595618 Fax: +39 06.2020.687
UCW gratefully acknowledges the support provided by the United States Department of Labor for the development of the report. This report does not necessarily reflect the views or policies of the United States Department of Labor or of the Global Affairs Canada. The mention of trade names, commercial products and organizations does not imply endorsement by the United States Government.
Farm cooperatives, household vulnerability and agricultural child labour in Rwanda
Contents
1. Background ......................................................................................................................... 1
2. Profile of agricultural child labour ...................................................................................... 3
3. Review of agricultural cooperatives in Rwanda .................................................................. 9
4. Empirical analysis of impact of agricultural cooperatives on farm child labour ............... 13
5. Conclusion ......................................................................................................................... 19
APPENDIX. Additional statistical tables ................................................................................... 20
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Overcoming the challenge of child labour will be critical to progress towards
the Sustainable Development Goals. The International Labour Organisation
(ILO) estimates that there were still some 120 million children aged 5-14 years
at work worldwide in 2012 and more than 47 million children aged 15-17 years
in hazardous work. In Rwanda, despite significant progress, a total of almost
180,000 children aged 7-15 years were in child labour in 2011, the latest year
for which data are available. By far the largest share of working children, both
globally and in Rwanda, are found in the agriculture sector.
Evidence shows that child labour is driven partially by household
vulnerabilities associated with poverty, risk and shocks, as these contingencies
can force households to resort to child labour as a coping strategy. Economic
vulnerability can be especially acute for agricultural households, as they face
uncertain crop prices and crop yields, and are susceptible to climate-related
shocks such as drought and flooding, all potentially impacting on their decisions
relating to child labour.1
Agricultural cooperatives offer an important means of reducing the
vulnerability of farm households. While the specific operational parameters
differ somewhat from cooperative to cooperative, most involve some form of
agronomic support, access to improved storage, processing and marketing, and
more stable pricing for agricultural outputs. Taken together, agricultural
cooperatives can play a vital role in increasing yields, reducing income volatility
and improving farm livelihoods. Their relevance to child labour is potentially
very significant – improved farm livelihoods mean less need to rely children’s
labour to make ends meet. As cooperatives are also often associated with more
efficient and advanced farming methods, they can also reduce the demand for
the manual labour provided by children on the farm.
1 See, for example, Kathleen Beegle, Rajeev H. Dehejia, Roberta Gatti “Child labor and agricultural shocks”, Journal of Development Economics 81 (2006) 80– 96.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
The current study addresses cooperatives, household vulnerability and child
labour in Rwanda. It specifically assesses the potential of agricultural
cooperatives as a strategy for addressing farm child labour in the country. The
study forms part of broader efforts to inform the operationalisation of the 2013
National Policy on the Elimination of Child Labour; the National Policy
recognises the complexity of the child labour phenomenon and the need to
mainstream it into sectoral strategies, plans, and budgets. The research builds
on previous collaboration with the Government of Rwanda, and in particular
with the Ministry of Public Service and Labour (MIFOTRA) and with the National
Institute of Statistics Rwanda (NISR).
The study is structured as follows. Following this introduction, Section 2
profiles child labour generally, and agricultural child labour in particular, in
Rwanda. Section 3 reviews agricultural cooperatives in Rwanda. Section 4 then
looks at links between agricultural child labour and involvement in agricultural
cooperatives. Section 5 concludes.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Child labour remains an important concern in Rwanda. Statistics from the
Integrated Household Living Conditions Survey (EICV III, 2010-2011) indicate
that six percent of children aged 6-15 years, 180,000 children in absolute terms,
are in child labour.2 This share rises to almost 14 percent among older, 12-15
year-old, children.
As reported in Table 1, child labour is about one-third higher in rural
compared to urban areas (six percent versus four percent), owing to the on-
going importance of children’s labour in the agricultural sector, as discussed in
more detail below. There are also large differences in terms of child labour
incidence across provinces. As also reported in Table 1, child labour in the
Northern province (12 percent) is almost four times higher than child labour in
Kigali city (three percent).
Table 1. Child labour, age group 6-15 years, percentage
(a) Percentage
Age
Sex Residence Province
Total
Male Female Urban Rural
Northern
province
Western
province
Southern
province
Eastern
province
Kigali
city
Total 6-15 6.2 6.0 4.2 6.4 12.3 5.7 5 3.6 3.3 6.1
6-11 1.9 1.9 1.0 2.0 5.3 1.9 1.3 0.3 0.4 1.9
12-15 13.7 13.3 9.5 14.1 24.0 12.6 11.5 9.5 8.5 13.5
2 Law No. 13/2009 regulating labour in Rwanda provides the primary legal framework regarding child labour in the country. Specifically, child labour provisions are contained in Title I, Chapter 3, Section 1 (Child labour prohibition) and in Title III, Chapter 4, Section 1 (Prohibited child labour and its nature) of Law 13/2009. The law prohibits employment of a child in any company, even as apprentice, before the age of 16 years (article 4), and sets specific conditions for the employment of 16-17 year-olds (articles 5 and 6). 68. It should be stressed that these child labour estimates are lower bound estimates, as they do not include involvement in what ILO terms “worst forms of child labour other than hazardous work”, which were beyond the scope of EICV III (2010-2011).
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Table 1.Cont’d
(b) Number
Age
Sex Residence Province
Total
Male Female Urban Rural
Northern
province
Western
province
Southern
province
Eastern
province
Kigali
city
Total 6-15 89,916 88,905 15,812 163,008 69,340 41,234 34,053 26,466 7,729 178,820
6-11 17,355 17,716 2,309 32,761 18,525 9,058 5,475 1,356 656 35,070
12-15 72,561 71,189 13,503 130,247 50,815 32,176 28,578 25,109 7,073 143,750
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
Rwanda has nonetheless seen important progress against child labour and is
one of the better performers in this regard in the Sub-Saharan Africa region. A
comparison of the results of the third Integrated Household Living Conditions
Survey fielded in 2010-2011 with those from the 2005 round of the same survey
shows that child labour declined considerably in Rwanda over the period
between the survey dates. As reported in Figure 1, the share of 6-15 year-olds
in child labour fell from 7.7 percent in 2005 to 6.1 percent in 2011, a reduction
of about one-fifth. In absolute terms, those in child labour dropped from
196,000 to 179,000. These reductions have occurred alongside significant
progress in reducing poverty and raising school enrolment rates; a detailed
discussion of the factors underlying the decline in child labour, however, is
beyond the scope of the current Report.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Figure 1. Child labour declined considerably over the period from 2005 to 2011
Changes in the number and percentage of children in child labour, 2005 and 2011
Source: UCW calculations based on Rwanda, EICV II (2005) and EICV III (2010-2011).
The overall decline in the number of child labourers was accounted for almost
entirely by the agriculture sector. Agricultural child labourers fell from 146,000
to 121,000 children over the period from 2005 to 2011, while numbers of
children in services and manufacturing actually increased over the same period.
By far the largest share of the remaining population of child labourers is
nonetheless found in agriculture and further progress towards child labour
elimination in Rwanda will therefore require a special focus on this sector. As
reported in Figure 2, 69 percent of all child labourers are found in agriculture,
rising to 72 percent in rural areas. The remaining child labourers are found in
services (24 percent and in industry (seven percent).
8.0
7.5 7.
7
6.2
6.0 6.1
99,8
92
95,9
80
195,
873
89,9
16
88,9
05
178,
821
0
25000
50000
75000
100000
125000
150000
175000
200000
225000
0
1
2
3
4
5
6
7
8
9
Male Female Total Male Female Total
num
ber
perc
ent
2005
2011
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Figure 2. Child labourers are concentrated in agricultural work
Sectoral distribution (%) of children in employment, age group 6-15 years, by sex and residence.
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
Children in agricultural child labour, in turn, are concentrated primarily in
non-waged work for their own families. Three quarters of agricultural child
labourers are in non-waged family work, while 19 percent are waged work and
six percent are in self-employment arrangements. This information enables a
further refinement in the targeting of efforts against child labour moving
Figure 3. Children’s agricultural work typically takes place within the family
Distribution (%) of children’s agricultural work by modality of employment, children aged 6-15 years
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
68.8 69.9
38.4
72.1
0.0
69.4
6.6 6.4
3.7
6.76.5
24.6 23.7
57.9
21.1 24.2
Male Female Urban Rural TOTAL
Sex . Residence . .
0
10
20
30
40
50
60
70
80
90
100
per
cen
t
Services
Industry
Agriculture
Wage agricultural work, 19.0
Agricultural work for self, 5.7
Non-wage family agricultural work, 75.0
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
forward – most child labourers are in agriculture and most agricultural child
labourers work for their own families.
Agricultural child labour is also a special concern in Rwanda because it
constitutes a much bigger barrier to school attendance than other forms of
child labour. This point is illustrated in Figure 4, which reports school
attendance rates for agricultural child labourers, other child labourers and
children not in child labour. As shown, agricultural child labourers are almost 10
percentage points less likely to attend school than other child labourers and
almost 20 percentage points less likely to be in school than children who are
not child labourers. Agricultural child labour, therefore, is not just an obstacle
to the achievement of child labour elimination and also to realizing universal
basic enrolment.
Figure 4. Agricultural child labour is the form of child labour that interferes most with children’s ability to attend school
School attendance (%), children aged 7-15 years, agricultural child labourers, other child labourers and children not in child labour
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
The families of agricultural child labourers are more likely to be low-income
than the families of other categories of child labourers. Forty percent of
agricultural households with child labourers are found in the lowest two income
quintiles, while the same statistic for households with other categories of child
labourers is just 10 percent. While these simple correlations should be
interpreted with caution, they suggest that poverty plays a larger relative role
68.7
77.8
87.0
0
10
20
30
40
50
60
70
80
90
Agricultural child labourers Other child labourers Children not in child labour
per
cen
t
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
in agricultural child labour, and, following from this, that policies aimed at
reducing economic vulnerability could be relevant to reducing agricultural child
labour. Agricultural cooperatives are one important option in this regard, as will
be discussed in more detail in the subsequent section.
Figure 5. 12. The families of agricultural child labourers are more likely to be low-income than the families of other categories of child labourers
Household income quintile distribution (%), children aged 7-15 years, agricultural and non-agricultural households with working children
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
4.5
19.15.4
20.7
9.5
21.7
15.4
22.9
65.2
15.6
0
10
20
30
40
50
60
70
80
90
100
Non-agricultural households with workingchildren
Agricultural households with workingchildren
per
cen
t
Highest incomequintile
4th
3rd
2nd
Lowest income quintile
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
The cooperative movement in Rwanda gained renewed importance in the
aftermath of the 1994 genocide as a vehicle for national reconciliation and
development.3 In order to establish a legal and administrative framework for
cooperatives, the Government set up a Taskforce on Cooperative Promotion in
2005, since transformed into the Rwanda Cooperative Agency. A National
Policy on Promotion of Cooperatives was released in 2006; the document lays
out “policy, strategies and action plan aimed at orienting and ensuring that
cooperatives become a viable tool for social-economic development in
Rwanda.”4 A series of laws were also promulgated dealing with the organisation
and functioning of cooperative organizations in Rwanda.5
The number of registered cooperatives in the operating in the country has
grown rapidly since the release of the National Policy. There were
approximately 2,500 registered cooperatives in 2008, while most recent figures
from the Rwanda Cooperative Agency indicate a total of 7,464 registered
primary cooperatives, alongside 479 savings and credit co-operatives
(SACCOs),6 131 unions and 14 federations.7 Numbers of unregistered
cooperatives are also substantial, although estimates in this regard vary
considerably. Many of the latter are “informal groups with a cooperative
3 For a review of the cooperation movement prior to 1994, see, inter alia, Herman Musahara (2012). “Perspectives on Cooperatives with Reference to Rwanda.” Paper presented at the regional conference on “Perspectives for cooperatives in Eastern Africa”, October 2012, in Kampala/Uganda. 4 National Policy on Promotion of Cooperatives, page 1. (http://www.rca.gov.rw/IMG/pdf/POLICY_DOCUMENT__FINAL__March_2006.pdf). 5 These include Law No. 50/2007 OF 18/09/2007 Providing FOR THE Establishment, Organisation and Functioning of Cooperative Organizations in Rwanda and Law No. 16/2008 OF 11/06/2008 Establishing Rwanda Cooperative Agency (RCA) and Determining Responsibilities, Organisation and Functioning. 6 A savings and credit co-operatives (SACCOs) is a type of co-operative whose objective is to pool savings for the members and in turn provide them with credit facilities. 7 Rwanda Cooperative Agency (http://www.rca.gov.rw/wemis/registration/all.php).
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
purpose” that do not yet meet all the requirements for recognition as
cooperative societies.8
The agriculture sector accounts for by far the largest share of cooperatives
and cooperative-like structures. The original Government baseline survey of
cooperatives undertaken in 2005 found that the sector accounted 69 percent
of all cooperative organisations (registered and unregistered), while credit-
related cooperatives accounted for 13 percent, handicrafts for five percent,
commerce and services each for four percent and fishing and construction each
for less than one percent.9
The cooperative movement has diversified considerably since the time of
the 2005 survey, although the agriculture sector continues to predominate.
Consumer cooperatives, producer, marketing, credit and saving, housing,
handicrafts and multipurpose are among the other forms of cooperatives cited
in the laws and identifiable in the country. Credit and saving cooperatives in
particular have gained in importance as a means of raising capital among low-
income groups including farmers.10
There has also been a shift over time in terms of the organisational structure
of the cooperative movement. Most of the multifunctional cooperatives and
cooperative unions covering specific geographical areas established during an
earlier period have ceased to exist, owing primarily to a lack of shared interest
among their members.11 In their place, there has been a trend towards vertical
structuring by product category. This trend is most visible in the agricultural
sector, and, within the sector, in rice and tea-growing. Rice growers have
8 ILO/Patrick Develtere, Ignace Pollet, and Fredrick O. Wanyama (2008). Cooperating out of Poverty. Chapter 11: Jump-starting the Rwandan cooperative movement. Geneva, ILO, 2008. 9 Rwanda Ministry of Trade and Industry, 2006, as cited in The hope for rural transformation: A rejuvenating cooperative movement in Rwanda, Coop AFRICA Working Paper No.12, Espérance Mukarugwiza, ILO, 2010. 10 Herman Musahara (2012). “Perspectives on Cooperatives with Reference to Rwanda.” Paper presented at the regional conference on “Perspectives for cooperatives in Eastern Africa”, October 2012, in Kampala/Uganda. 11 ILO/Patrick Develtere, Ignace Pollet, and Fredrick O. Wanyama (2008). Cooperating out of Poverty. Chapter 11: Jump-starting the Rwandan cooperative movement. Geneva, ILO, 2008.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
succeeded in forming grassroots cooperatives that are grouped into
cooperative unions that in turn form a national federation. Similarly, among tea-
growers, a national federation exists (Ferwathé) that is responsible for all tea
producers who belonging to cooperatives. The over 400,000 coffee growers in
Rwanda are less organised, although there have been important success stories
among the coffee cooperatives that are in operation. In the potato, wheat,
maize and cassava markets, the organising process is accelerating, while in
other areas such as housing, fishing and livestock husbandry, the existing
voluntary associations and cooperatives are only very loosely structured.12
Reliable recent statistics on the scale and coverage of the cooperative
movement are scarce. One exhaustive survey in Ruhengeri (one of the eleven
original provinces) in 2003, however, suggests coverage is substantial: the
survey indicated a 4,869 recorded organizations with 231,456 members,
corresponding to almost half of the province’s adult population at the time.13
Within the agriculture sector, one measure of coverage is the share of farms
selling their output to agricultural cooperatives. Data for this measure from the
Integrated Household Living Conditions Survey (EICV III, 2010-2011) indicate
that about eight percent family farms use cooperatives as the principal
outlet/market for their agricultural production.
Rwandan cooperatives appear to be relevant to alleviating household
vulnerability and improving livelihoods, although the absence of any formal
impact studies means conclusions in this regard must be drawn with caution.
The cooperatives, particularly in the informal economy, help small farmers by
pooling their bargaining power and providing them with more stable crop
pricing. There is anecdotal evidence of how cooperative membership has
improved (and smoothed) incomes, enabling members, for instance, to build
12 ILO/Patrick Develtere, Ignace Pollet, and Fredrick O. Wanyama (2008). Cooperating out of Poverty. Chapter 11: Jump-starting the Rwandan cooperative movement. Geneva, ILO, 2008. 13 Care International (2003), as cited in ILO/Patrick Develtere, Ignace Pollet, and Fredrick O. Wanyama (2008). Cooperating out of Poverty. Chapter 11: Jump-starting the Rwandan cooperative movement. Geneva, ILO, 2008.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
houses, pay their children’s school fees, produce food for their families and
improve their clothing. In one specific example, women weavers visited at a
cooperative in Bugesera report their husbands’ appreciation of their
contribution to meeting family needs.14 Cooperatives have also been used to
tackle social exclusion, helping groups such as demobilized soldiers, people
living with HIV/AIDS, genocide survivors (especially widows and orphans), ex-
prisoners, street children and commercial sex workers, among others, in
starting income-generating activities.15
Cooperatives are also linked more directly with household social protection
and risk mitigation, particularly in the field of health. Cooperatives have made
it easier for their members to pay premiums for health insurance schemes,
referred to as mutuelle de santé. In the tea sector, for instance, members can
apply for loans to cover the annual contributions to the mutuelle de santé,
which they then gradually pay back to the cooperative. Cooperatives also
facilitate the development of mutual assistance schemes, for example, during
bereavement or the extraordinary costs associated with social functions such
as wedding ceremonies. HIV/AIDS awareness and protection campaigns are
other areas where cooperatives have played an important role, providing a key
entry point for sensitization and training efforts. 16
14 The hope for rural transformation: A rejuvenating cooperative movement in Rwanda, Coop AFRICA Working Paper No.12, Espérance Mukarugwiza, ILO, 2010. 15 Ibid, 2010. 16 Ibid, 2010.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
The preceding discussion makes clear that agricultural cooperatives can play
an important role in reducing the economic vulnerability of farm households.
This in turn raises the question of their potential relevance in the context of
agricultural child labour, as children’s labour in many contexts can be used as a
household coping mechanism in the face of economic vulnerability. Agricultural
cooperatives, in other words, could, at least in theory, help reduce agricultural
child labour by reducing the economic vulnerability that drives it. To the extent
that agricultural cooperatives also help introduce improved and more advanced
farming methods, they can also reduce the need for the manual labour provided
by children on the farm. It is also not possible, however, to exclude a priori the
possibility that cooperatives increase the demand for children’s labour by
encouraging expanded agricultural production.
In this section, the actual relationship between agricultural child labour and
farm cooperatives is explored making use of data from Integrated Household
Living Conditions Survey (EICV III, 2010-2011). The survey collected information
on whether or not the principal outlet/market of agricultural production of a
farm was a cooperative, which is used as a proxy for membership in an
agricultural cooperative.17 This information, coupled with information from the
same survey on involvement in agricultural child labour and schooling, permits
an interesting initial look into how agricultural child labour and school
attendance are correlated with cooperative membership.
A simple comparison of family farms selling to cooperatives and other family
farms indicates that child labour is much more limited on the former. As
reported in Figure 6, child labour is one-third lower on family farms selling to
cooperatives (six percent versus four percent); the difference is especially
17 The survey also collected information on whether or not a household member was working in a business that was part of a cooperative or of a business association, which serves as a proxy for membership in a non-agricultural cooperative.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
pronounced among females (44 percent lower) and among younger children
(42 percent lower).
Figure 6. Involvement in child labour is significantly lower on family farms selling to cooperatives
Percentage difference between child labour on farms selling to cooperatives and child labour on other farms, by age range and sex(a)
Notes: (a) Difference in child labour between other farms and farms selling to cooperatives, expressed as a percentage of child labour on other farms.
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
The time intensity of child labour is also lower on farms selling to
cooperatives, as reported in Figure 7. Overall, children working on farms selling
to cooperatives put in almost 12 percent fewer hours each week; again the
differences are especially large for females (22 percent) and younger children
(48 percent). Simple descriptive comparisons, therefore, indicate that children
on farm cooperatives work less at both the extensive and intensive margins
than their peers from farm families that are not cooperative members.
-42.1
-32.4
-22.2
-43.5
-33.3
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
5
6-11 years 12-15 years Male Female TOTAL
Age range . Sex .
per
cen
t
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Figure 7. The time intensity of child labour is also significantly lower on family farms selling to cooperatives
Percentage difference in average weekly working hours between children working on farms selling to cooperatives and children working on other farms, by age range and sex(a)
Notes: (a) Difference in working hours between other farms and farms selling to cooperatives, expressed as a percentage of working hours on other farms.
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
The negative correlation between cooperative membership and child labour
holds even when controlling for possible confounding individual and household
factors through use of regression analysis. The variables and descriptive results
in this regard are reported in Appendix Table A1.
Three different econometric specifications are used. The first two
specifications look at the impact of farm cooperative membership on
agricultural child labour and schooling controlling for various background
characteristics and two different income (welfare) measures as a robustness
check. The first controls for the distribution of the household income quintile
and the second for household poverty status (extreme-poor, poor, non-poor).
The third specification looks at the impact non-farm cooperative membership
on child labour outside the agriculture sector.18
The estimation results relating to membership in an agricultural
cooperative, presented in Tables 2 and 3 below, indicate that agricultural
18 The survey also collected information on whether or not a household member was working in a business that was part of a cooperative or of a business association; this information is used as a proxy for membership in a non-agricultural cooperative.
-47.7
-10.8
-4.5
-22.1
-11.6
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
6-11 years 12-15 years Male Female TOTAL
Age range . Sex .
per
cen
t
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
cooperative membership significantly reduces the probability of child labour,
consistent with the descriptive results presented above. In both specifications,
children from farm households belonging to agricultural cooperatives are about
two percentage points less likely to be child labourers than children from other
farm households. While the agricultural cooperative membership does not
increase the likelihood of school attendance, it does increase the likelihood that
children are able to attend school without also being burdened by work (by
about two percentage points in both specifications).
Table 2. Marginal effects after bivariate probit estimation, children aged 6-15,(a) specification 1
In child labour only In school only In child labour and
school Neither in child labour
nor in school
dy/dx SE dy/dx SE dy/dx SE dy/dx SE
Age (years) -0.038*** 0.002 0.273*** 0.007 0.034*** 0.005 -0.270*** 0.005
Age squared 0.002*** 0.000 -0.013*** 0.000 -0.001*** 0.000 0.012*** 0.000
Female -0.003*** 0.001 0.022*** 0.005 0.001 0.003 -0.020*** 0.004
household size -0.001*** 0.000 0.005*** 0.001 -0.004*** 0.001 0.000 0.001
Household head education(b)
Primary -0.004*** 0.001 0.022*** 0.006 -0.002 0.003 -0.017*** 0.005
Post primary -0.010*** 0.002 0.074*** 0.012 -0.004 0.007 -0.059*** 0.009
Secondary -0.008*** 0.002 0.049*** 0.013 -0.011* 0.006 -0.031*** 0.010
University -0.012*** 0.002 0.083*** 0.018 -0.013 0.009 -0.058*** 0.014
Agriculture cooperative -0.004** 0.002 0.015* 0.009 -0.017*** 0.005 0.005 0.007
Household Income quintile(c)
2nd quintile -0.004*** 0.001 0.033*** 0.008 0.012*** 0.004 -0.041*** 0.006
3rd quintile -0.009*** 0.001 0.060*** 0.008 0.001 0.004 -0.052*** 0.006
4th quintile -0.009*** 0.001 0.064*** 0.008 -0.001 0.004 -0.053*** 0.006
5th quintile -0.010*** 0.002 0.075*** 0.009 0.008 0.005 -0.072*** 0.007
Area of residence(d)
Rural 0.004*** 0.001 -0.017** 0.008 0.016*** 0.004 -0.003 0.006
Notes: (a) *** p<0.01, ** p<0.05, * p<0.; (b) Household head education: reference category "no education"; (c) Household income quintile: reference category "bottom quintile"; and (d) Area of residence: reference category "urban area"
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Table 3. Marginal effects after bivariate probit estimation, children aged 6-15,(a) specification 2
In child labour only In school only In child labour and
school Neither in child labour
nor in school
dy/dx SE dy/dx SE dy/dx SE dy/dx SE
Age (years) -0.038*** 0.002 0.273*** 0.007 0.034*** 0.005 -0.269*** 0.005
Age squared 0.002*** 0.000 -0.013*** 0.000 -0.001*** 0.000 0.012*** 0.000
Female -0.003*** 0.001 0.022*** 0.005 0.001 0.003 -0.020*** 0.004
Household size -0.001*** 0.000 0.005*** 0.001 -0.004*** 0.001 0.000 0.001
Household head education(b)
Primary -0.004*** 0.001 0.023*** 0.006 -0.002 0.003 -0.018*** 0.005
Post primary -0.011*** 0.002 0.076*** 0.012 -0.003 0.007 -0.063*** 0.009
Secondary -0.008*** 0.002 0.054*** 0.012 -0.008 0.006 -0.038*** 0.009
University -0.012*** 0.002 0.090*** 0.017 -0.009 0.010 -0.069*** 0.012
Agriculture cooperative -0.004** 0.002 0.016* 0.009 -0.017*** 0.005 0.005 0.007
Poverty status(c)
Poor -0.004*** 0.001 0.030*** 0.007 0.007* 0.004 -0.033*** 0.005
Non-poor -0.009*** 0.001 0.059*** 0.006 -0.001 0.003 -0.049*** 0.005
Area of residence(d)
Rural 0.004*** 0.001 -0.020** 0.008 0.015*** 0.004 0.000 0.006
Notes: (a) *** p<0.01, ** p<0.05, * p<0.; (b) Household head education: reference category "no education"; (c) Poverty status: reference category "extreme poor" and (d) Area of residence: reference category "urban area"
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
The results of the third specification, reported in Table 4, suggest that the
child labour impact of cooperative membership does not extend to
cooperatives outside the agriculture sector. Membership in non-agricultural
cooperatives has no significant impact on either child labour or schooling.
Table 4. Marginal effects after bivariate probit estimation, children aged 6-15,(a) specification 3
In child labour only In school only In child labour and
school Neither in child labour
nor in school
dy/dx SE dy/dx SE dy/dx SE dy/dx SE
Age (years) -0.038*** 0.002 0.273*** 0.007 0.034*** 0.005 -0.270*** 0.005
Age squared 0.002*** 0.000 -0.013*** 0.000 -0.001*** 0.000 0.012*** 0.000
Female -0.003*** 0.001 0.022*** 0.005 0.001 0.003 -0.020*** 0.004
household size -0.001*** 0.000 0.005*** 0.001 -0.004*** 0.001 0.000 0.001
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Table 4.Cont’d
In child labour only In school only In child labour and
school Neither in child labour
nor in school
Household head education(b)
Primary -0.004*** 0.001 0.022*** 0.006 -0.002 0.003 -0.017*** 0.005
Post primary -0.010*** 0.002 0.074*** 0.012 -0.004 0.007 -0.059*** 0.009
Secondary -0.008*** 0.002 0.049*** 0.013 -0.011* 0.006 -0.030*** 0.010
University -0.012*** 0.002 0.082*** 0.018 -0.012 0.009 -0.058*** 0.014
Non-agriculture cooperative -0.002 0.003 0.016 0.019 -0.001 0.010 -0.013 0.014
Household Income quintile(c)
2nd quintile -0.004*** 0.001 0.033*** 0.008 0.011*** 0.004 -0.041*** 0.006
3rd quintile -0.009*** 0.001 0.061*** 0.008 0.000 0.004 -0.052*** 0.006
4th quintile -0.010*** 0.001 0.065*** 0.008 -0.002 0.004 -0.053*** 0.006
5th quintile -0.011*** 0.002 0.074*** 0.009 0.007 0.005 -0.071*** 0.007
Area of residence(d)
Rural 0.004*** 0.001 -0.017** 0.008 0.016*** 0.004 -0.003 0.006
Notes: (a) *** p<0.01, ** p<0.05, * p<0.; (b) Household head education: reference category "no education"; (c) Household income quintile: reference category "bottom quintile"; and (d) Area of residence: reference category "urban area"
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA
Remaining child labourers in Rwanda are concentrated in the agricultural
sector and further progress towards child labour elimination will therefore
require a special focus on this sector. This is the first study in Rwanda to assess
the potential role of the cooperative movement in addressing agricultural child
labour.
Using unique data on cooperatives from the Integrated Household Living
Conditions Survey (EICV III, 2010-2011) the results indicate a significant
negative correlation between cooperative membership and involvement in
agricultural child labour. Controlling for individual and household
characteristics, including income, children from farm households belonging to
agricultural cooperatives are about one-third less likely to be child labourers
than children from other farm households. Agricultural cooperative
membership also makes it significantly more likely that children are able to
attend school without having to shoulder the additional burden of farm labour.
It is interesting to note that these results do not extend to membership in other
types of non-farm cooperatives.
The preliminary results presented in this study hint at the potential of the
cooperative movement in Rwanda as a means of addressing agricultural child
labour. More research is needed to confirm and explain this result, but it is likely
that cooperative membership helps reduce the economic vulnerability of
agricultural households, in turn reducing their reliance on their children’s
labour.
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA: POLICY APPRAISAL
Table A1.
Variable Obs Mean Std. Dev. Min Max
Child and household characteristics
child labour 18599 6.10 23.94 0 1
school attendance 18611 86.03 34.67 0 1
Age (years) 18611 10.30 2.87 6 15
Age squared 18611 114.35 60.50 36 225
Male 18611 0.50 0.50 0 1
Female 18611 0.50 0.50 0 1
household size 18611 6.44 2.09 1 22
Household head education
No education 18605 0.26 0.44 0 1
Primary 18605 0.61 0.49 0 1
Post primary 18605 0.05 0.21 0 1
Secondary 18605 0.06 0.23 0 1
University 18605 0.03 0.16 0 1
Cooperative coop_agric 18611 0.08 0.28 0 1
Non-agriculture cooperative 18611 0.02 0.14 0 1
Poverty status
Extremely poor 18611 0.29 0.45 0 1
Poor 18611 0.22 0.41 0 1
Non-poor 18611 0.49 0.50 0 1
Household income quintile
Bottom quintile 18611 0.25 0.43 0 1
2nd quintile 18611 0.22 0.41 0 1
3rd quintile 18611 0.20 0.40 0 1
4th quintile 18611 0.17 0.38 0 1
Top quintile 18611 0.17 0.37 0 1
Residence Urban 18611 0.13 0.33 0 1
Rural 18611 0.87 0.33 0 1
FARM COOPERATIVES, HOUSEHOLD VULNERABILITY AND AGRICULTURAL CHILD LABOUR IN RWANDA: POLICY APPRAISAL
Table A2.
reg1 reg2 reg3
Child labour School attendance Child labour School attendance Child labour School attendance
Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE Coef. SE
Age (years) -0.034 0.058 1.922*** 0.040 -0.035 0.058 1.914*** 0.040 -0.032 0.058 1.922*** 0.040
Age square 0.011*** 0.003 -0.085*** 0.002 0.011*** 0.003 -0.084*** 0.002 0.011*** 0.003 -0.085*** 0.002
Female -0.027 0.032 0.144*** 0.027 -0.027 0.032 0.143*** 0.027 -0.024 0.032 0.144*** 0.027
Household size -0.054*** 0.008 0.006 0.007 -0.054*** 0.008 0.006 0.007 -0.056*** 0.008 0.006 0.007
Household head education
Primary -0.053 0.037 0.121*** 0.031 -0.050 0.036 0.127*** 0.031 -0.056 0.037 0.121*** 0.031
Post primary -0.147* 0.089 0.481*** 0.081 -0.136 0.088 0.504*** 0.080 -0.148* 0.089 0.481*** 0.081
Secondary -0.199** 0.088 0.241*** 0.074 -0.167* 0.086 0.295*** 0.072 -0.191** 0.088 0.239*** 0.074
University -0.274* 0.140 0.482*** 0.127 -0.227* 0.135 0.578*** 0.123 -0.263* 0.140 0.481*** 0.127
Agriculture cooperative -0.204*** 0.063 -0.009 0.049 -0.207*** 0.063 -0.007 0.049
Household income quintile
2nd quintile 0.070 0.045 0.251*** 0.037 0.067 0.045 0.251*** 0.037
3rd quintile -0.079 0.049 0.357*** 0.040 -0.086* 0.049 0.356*** 0.040
4th quintile -0.112** 0.052 0.369*** 0.042 -0.118** 0.052 0.368*** 0.042
Highest quintile -0.028 0.059 0.508*** 0.051 -0.031 0.059 0.506*** 0.051
Area of residence
Rural 0.226*** 0.057 -0.005 0.045 0.212*** 0.056 -0.029 0.044 0.219*** 0.057 -0.005 0.044
Poverty status
Poor 0.026 0.044 0.212*** 0.036
Non-poor -0.102*** 0.039 0.350*** 0.032
Non Farm cooperative -0.031 0.121 0.094 0.101
_cons -2.463*** 0.327 -9.222*** 0.204 -2.430*** 0.326 -9.118*** 0.202 -2.466*** 0.327 -9.224*** 0.204
/athrho -0.591*** 0.028 -0.588*** 0.028 -0.591*** 0.028
Notes: (a) *** p<0.01, ** p<0.05, * p<0.; (b) Household head education: reference category "no education"; (c) Household income quintile: reference category "bottom quintile"; and (d) Area of residence: reference category "urban area"
Source: UCW calculations based on Rwanda, EICV III, 2010-2011.